搜索“NVIDIA”

he question was how much of that advantage the rest of the industry could actually get. Could you take an open-weight model, serve it through a familiar inference engine, and beat NVIDIA on the economics that matter to your business? Today, we are publishing the first third-party inference results for TPUv7 Ironwood on InferenceX Official Preview. In our apples-to-apples comparisons against B200/B300, Ironwood delivers up to 50% better performance per dollar. Its advantage extends across much of

匹配字段正文

icians had budgeted years for that ran to 13M lines of code. Jensen Huang said “AGI has arrived” in a post about OpenAI’s GPT-6 Astra, revealing that the model was trained on 100K Nvidia GPUs with 400K more “coming online next.” The U.S. and China are reportedly lining up AI safety talks for mid-September, with topics including AI cyberattack monitoring and model distillation. Los Angeles public schools shut off student access to AI on district devices for the school year, coming on the heels of

匹配字段正文

ven beforehand); they would tell you to be scaling pilled. Well, we didn’t actually see that many of them throwing away everything to get more compute, more data, and investing in NVIDIA call options. Similarly, people talk about being AGI pilled. Well, if that’s your belief, that AGI will hit in the next year or two with the current transformer paradigm, why aren’t you buying as much land as possible, or folding your little neolab startup and begging one of the AGI-leading labs to let you in? I

匹配字段正文

You know, sometimes it’s kind of hard to explain the “circular” part of circular financing to people, in the sense that some of the agreements are kind of clunky. NVIDIA funds OpenAI, who then spends that money to rent back NVIDIA GPUs from Microsoft, Google, Amazon, or CoreWeave, and then that money is used…to buy servers from Taiwanese ODMs (original design manufacturers) that build their servers, who then buy GPUs from NVIDIA to put in them. The reason it’s clunky is that people will, even if

匹配字段正文

ries at the G20 Innovation Ministerial. Too many AI auditors come from a similar POV. Anthropic and OpenAI’s professional power-users account for most of their enterprise revenue. Nvidia and Hugging Face announce the chipmaker’s $12.9 billion acquisition agreement. OpenAI pulls back the curtain on its new Astra model. Pablo Torre’s investigation into the LA Clippers results in historic NBA sanctions and a Pulitzer for his team. The Trump administration weighs in on behalf of OpenAI in its legal

匹配字段正文

lagship, top-rated model 🗣️ Realtime-TTS-2 - Inworld's top-ranked voice model, now generally available 🌎 Atlas - World Labs’ new SOTA world model in early access Sponsored Listing Nvidia announced the acquisition of Hugging Face for $12.9B, with CEO Jensen Huang saying the platform will remain open to every cloud and chip. Axiom Math published a paper breaking a decade-old record by proving prime numbers never stop appearing within 212 of each other, only for OpenAI's GPT-6 Astra to shrink that

匹配字段正文

性缺陷都意味着不可承受之重。 面对这四大趋势，Gupta提出了“AI原生设计”的战略构想：将AI深度融入从芯片、封装、电路板到系统的全生命周期开发流程。其核心是“更快的引擎速度、更智能的执行、可信的结果”三大支柱。 在技术实现上，Gupta详细介绍了西门子EDA构建的自主智能体系统。该系统依托2025年7月发布的Fuse基础模型及其构建的结构化数据本体，基于NVIDIA的OpenShell框架实现智能体互联，依托NeMo Gym完成智能体训练。整个系统采用五层架构设计：底层为Arm CPU与GPU驱动的加速计算引擎，之上依次为高级推理层、智能体层与编排层，顶层则是西门子自研EDA引擎。 实践案例一：Solido定制IC验证平台。这是西门子将自主智能体AI工作流落地的首个标杆。工程师只需以自然语言下达指令（例如“对所有单元进行特征化”），系统即可自主调度数千个单元并覆盖全PVT（工艺、电压、温度）角点，独立完成仿真运行、结果调试与错误修复等复杂流程，实现全程无人值守。该方案将特征化速度提升10倍、Token成本降低10倍。Solido团队拥有超过二十年深耕AI应用于EDA的经验，为这一突

匹配字段正文

rompt — it can watch a human do a job once, then run the same job on hardware with no fine-tuning and no new training data. The details: Skild (valued at $14B, backed by SoftBank, Nvidia, and Jeff Bezos) is building a universal “brain” designed to run everything from robot arms to humanoids. S1 handles jobs missing from its pretraining, including potting a plant and flipping pancakes, with some running for 10 minutes across dozens of steps. In the plant test, just 11 minutes passed between start

匹配字段正文

memory across the cluster. These abstractions are already implemented in various open-source libraries, which we’ll use in this work, mainly https://github.com/ai-dynamo/nixl from nvidia-dynamo team, that builds on top of https://github.com/openucx/ucx - Unified Communication X: a library that abstracts the transfer-level primitives, allowing NIXL to be somewhat transfer level agnostic. This also lets us to not worry about things as rkey, lkey and memory consistency models. Sky is the limit To s

匹配字段正文

ks It won't kill you, or humanity, to get off the AI upgrade treadmill New models arrive so fast, we already have more new tech than it's possible to understand in a career Cisco, Nvidia team up to build fatter, faster pluggable If you thought 1.6 Tbps was fast, just wait, Cisco's USFP targets 12.8 Tbps in a single optical pluggable TOP STORIES Anthropic asks users to stop being mean to Claude; '123456' password linked to Denmark data breach affecting 8.8M people; Microsoft and Anthropic play in

匹配字段正文

ler Marvell – are still humming along with their ups and downs, but Marvell does not really talk about these much anymore except as they are lumped into a single group. Like this: Nvidia, for what it is worth, has done the same thing, lumping everything that is not part of its datacenter business into a group called Edge Computing. In fact, this chart looks very much like Nvidia’s chart, although the numbers on the Y axis are a hell of a lot smaller and the exponential curve is not quite so stee

匹配字段正文

nds over a company worth around $4.6T, up from about $350B when he took the job in 2011 — Apple briefly became the second company ever to touch $5T on July 28, a day after passing Nvidia. Ternus’s first keynote is Sept. 9, covering new iPhone hardware and the redesigned Siri, built on Gemini models. Image source: Instagram The Rundown: Instagram is now making AI personas identify themselves: accounts featuring an AI-generated human must toggle on a disclosure label, or they’ll stop showing up in

匹配字段正文

e CEO of Rhoda AI, told me that the company has “literally over 100 different use cases” in manufacturing and warehousing they’re working on bringing to deployment. When I went to Nvidia GTC in March, several roboticists told me that the most impressive demo they saw was from a company called Generalist. Generalist showed a robot inserting and removing a phone from its box using two robotic arms. While the task itself is pretty difficult for current robots, what impressed people the most was how

匹配字段正文

ffort to do this back in 2022, four or five years ago. But at that time the large models weren’t that great at programming, so it didn’t really take off. Max and his colleagues at NVIDIA, Jim Fan , Yuke Zhu , and others noticed that these new models were getting really good. Simultaneously, the OpenClaw breakthrough happened. Suddenly, agents collaborating with other agents really took off. This opened up new terrain – it was exactly the missing link I’d been looking for. Because large foundatio

匹配字段正文

egration in ComfyUI. Follow our installation tutorials for Trellis.2 and Pixal3D. Licensing Trellis.2’s own code and weights are MIT-licensed, but its original pipeline depends on NVIDIA’s nvdiffrast (for mesh rasterization) and nvdiffrec (for Physically Based Rendering), both distributed under the NVIDIA Source Code License which restricts usage to non-commercial research and evaluation. In practice, a studio couldn’t ship assets from the reference pipeline without stepping into a legal gray zo

匹配字段正文

ks It won't kill you, or humanity, to get off the AI upgrade treadmill New models arrive so fast, we already have more new tech than it's possible to understand in a career Cisco, Nvidia team up to build fatter, faster pluggable If you thought 1.6 Tbps was fast, just wait, Cisco's USFP targets 12.8 Tbps in a single optical pluggable TOP STORIES Anthropic asks users to stop being mean to Claude; '123456' password linked to Denmark data breach affecting 8.8M people; Microsoft and Anthropic play in

匹配字段正文

matters: The duck is cute, but the pitch underneath is that cheap programmable hardware is how developers get into physical AI. Pollen has already sold 10K+ Reachy Minis, and with Nvidia reportedly agreeing to buy Hugging Face for $12.9B, Microduck adds another accessible piece of hardware to its serious robotics push . Image source: Boston Dynamics The Rundown: ICE is preparing to spend up to $2M on Boston Dynamics’ Spot robot dogs, adding the four-legged machines to its enforcement arsenal for

匹配字段正文

mem设备可以跟踪来自主机的读所有权请求，并了解其自身的计算核心何时可以安全地修改数据。我不确定MX1是否为其RISC-V核心充分利用了这一能力，但三星表示该设备可以使用CXL内存语义在CPU和GPU之间共享其内存。希望这意味着该设备可以使用监听来与处理器内存子系统集成。 与LPDDR5X-PIM一样，MX1的吞吐量并不算高。作为参照，一块板载内存带宽相近的Nvidia GeForce GTX 1080拥有8.8 TFLOPS的FP32算力，而MX1仅能提供3 TFLOPS。但原始吞吐量并非重点。MX1提供的是一种缓解内存扩展部分弊端的方法。MX1侧的计算可以避开受限的主机接口，并避免穿越CXL链路带来的功耗开销。一个假设拥有8.8 TFLOPS算力的加速器，如果必须通过CXL为每个FLOP加载一个字节，其算力将被限制在仅64 GFLOPS。在同样场景下，MX1即使同样未命中缓存，也能超过200 GFLOPS。 MX1的近存计算方案展现出前景。CXL内存扩展器到其DRAM池的带宽往往高于CXL链路所能承载的带宽。常规主机访问会让其中很大一部分带宽闲置，如果主机无法从缓存中满足足够多的访

匹配字段正文

ks It won't kill you, or humanity, to get off the AI upgrade treadmill New models arrive so fast, we already have more new tech than it's possible to understand in a career Cisco, Nvidia team up to build fatter, faster pluggable If you thought 1.6 Tbps was fast, just wait, Cisco's USFP targets 12.8 Tbps in a single optical pluggable TOP STORIES Anthropic asks users to stop being mean to Claude; '123456' password linked to Denmark data breach affecting 8.8M people; Microsoft and Anthropic play in

匹配字段正文

(including Anthropic) warning that AI-enabled cyberattacks will become “far more widespread and sophisticated” in the coming months, calling for a coordinated cyber defense push. Nvidia is reportedly acquiring open model hub Hugging Face for $12.9B, with open-source interest continuing to surge on Chinese model gains. Google released Gemini Omni 1.1 Flash, an upgraded video model that adds 40-sec scene extensions and 4K upscaling, moving to the top spot on Arena’s text-to-image leaderboard. Thi

匹配字段正文